Search Results (32)

The bioaccessibility of bioactive compounds plays a major role in the nutritional value of foods, but there is a lack of systematic studies assessing the effect of the food matrix on bioaccessibility. Curcuminoids are phytochemicals extracted from Curcuma longa that have captured public attention due to claimed health benefits. The aim of this study is to develop a mathematical model to predict curcuminoid’s bioaccessibility in biscuits and custard based on different fibre type formulations. Bioaccessibilities for curcumin-enriched custards and biscuits were obtained through in vitro digestion, and physicochemical food properties were characterised. A strong correlation between macronutrient concentration and bioaccessibility was observed (p = 0.89) and chosen as a main explanatory variable in a Bayesian hierarchical linear regression model. Additionally, the patterns of food matrix effects on bioaccessibility were not the same in custards as in biscuits; for example, the hemicellulose content had a moderately strong positive correlation to bioaccessibility in biscuits (p = 0.66) which was non-significant in custards (p = 0.12). Using a Bayesian hierarchical approach to model these interactions resulted in an optimisation performance of r² = 0.97 and a leave-one-out cross-validation score (LOOCV) of r² = 0.93. This decision-support system could assist the food industry in optimising the formulation of novel food products and enable consumers to make more informed choices. Full article

Text classification is not only a prerequisite for natural language processing work, such as sentiment analysis and natural language reasoning, but is also of great significance for screening massive amounts of information in daily life. However, the performance of classification algorithms is always affected due to the diversity of language expressions, inaccurate semantic information, colloquial information, and many other problems. We identify three clues in this study, namely, core relevance information, semantic location associations, and the mining characteristics of deep and shallow networks for different information, to cope with these challenges. Two key insights about the text are revealed based on these three clues: key information relationship and word group inline relationship. We propose a novel attention feature fusion network, Attention Pyramid Transformer (APTrans), which is capable of learning the core semantic and location information from sentences using the above-mentioned two key insights. Specially, a hierarchical feature fusion module, Feature Fusion Connection (FFCon), is proposed to merge the semantic features of higher layers with positional features of lower layers. Thereafter, a Transformer-based XLNet network is used as the backbone to initially extract the long dependencies from statements. Comprehensive experiments show that APTrans can achieve leading results on the THUCNews Chinese dataset, AG News, and TREC-QA English dataset, outperforming most excellent pre-trained models. Furthermore, extended experiments are carried out on a self-built Chinese dataset theme analysis of teachers’ classroom corpus. We also provide visualization work, further proving that APTrans has good potential in text classification work. Full article

Effective lane detection technology plays an important role in the current autonomous driving system. Although deep learning models, with their intricate network designs, have proven highly capable of detecting lanes, there persist key areas requiring attention. Firstly, the symmetry inherent in visuals captured by forward-facing automotive cameras is an underexploited resource. Secondly, the vast potential of position information remains untapped, which can undermine detection precision. In response to these challenges, we propose FF-HPINet, a novel approach for lane detection. We introduce the Flipped Feature Extraction module, which models pixel pairwise relationships between the flipped feature and the original feature. This module allows us to capture symmetrical features and obtain high-level semantic feature maps from different receptive fields. Additionally, we design the Hierarchical Position Information Extraction module to meticulously mine the position information of the lanes, vastly improving target identification accuracy. Furthermore, the Deformable Context Extraction module is proposed to distill vital foreground elements and contextual nuances from the surrounding environment, yielding focused and contextually apt feature representations. Our approach achieves excellent performance with the F1 score of 97.00% on the TuSimple dataset and 76.84% on the CULane dataset. Full article

(1) Background: The effective selection of hyperspectral feature bands is pivotal in monitoring the nutritional status of intricate alpine grasslands on the Qinghai–Tibet Plateau. The traditional methods often employ hierarchical screening of multiple feature indicators, but their universal applicability suffers due to the use of a consistent methodology across diverse environmental contexts. To remedy this, a backward feature elimination (BFE) selection method has been proposed to assess indicator importance and stability. (2) Methods: As research indicators, the crude protein (CP) and chlorophyll (Chl) contents in degraded grasslands on the Qinghai–Tibet Plateau were selected. The BFE method was integrated with partial least squares regression (PLS), random forest (RF) regression, and tree-based regression (TBR) to develop CP and Chl inversion models. The study delved into the significance and consistency of the forage quality indicator bands. Subsequently, a path analysis framework (PLS-PM) was constructed to analyze the influence of grassland community indicators on Spec_Chl and Spec_CP. (3) Results: The implementation of the BFE method notably enhanced the prediction accuracy, with ΔR²_RF-Chl = 56% and ΔR²_RF-CP = 57%. Notably, spectral bands at 535 nm and 2091 nm emerged as pivotal for CP prediction, while vegetation indices like the PRI and mNDVI were critical for Chl estimation. The goodness of fit for the PLS-PM stood at 0.70, indicating the positive impact of environmental factors such as grassland cover on Spec_Chl and Spec_CP prediction (r_Chl = 0.73, r_CP = 0.39). Spec_Chl reflected information pertaining to photosynthetic nitrogen associated with photosynthesis (r = 0.80). (4) Disscusion: Among the applied model methods, the BFE+RF method is excellent in periodically discarding variables with the smallest absolute coefficient values. This variable screening method not only significantly reduces data dimensionality, but also gives the best balance between model accuracy and variables, making it possible to significantly improve model prediction accuracy. In the PLS-PM analysis, it was shown that different coverage and different community structures and functions affect the estimation of Spec_CP and Spec_Chl. In addition, Spec_Chl has a positive effect on the estimation of Spec_CP (r = 0.80), indicating that chlorophyll does reflect photosynthetic nitrogen information related to photosynthesis, but it is still difficult to obtain non-photosynthetic and compound nitrogen information. (5) Conclusions: The application of the BFE + RF method to monitoring the nutritional status of complex alpine grasslands demonstrates feasibility. The BFE filtration process, focusing on importance and stability, bolsters the system’s generalizability, resilience, and versatility. A key research avenue for enhancing the precision of CP monitoring lies in extracting non-photosynthetic nitrogen information. Full article

The burgeoning domain of medical imaging has witnessed a paradigm shift with the integration of AI, particularly deep learning, enhancing diagnostic precision and expediting the analysis of Computed Tomography (CT) images. This study introduces an innovative Multilayer Perceptron-driven model, DiagnosticMLP, which sidesteps the computational intensity of attention-based mechanisms, favoring a no-attention architecture that leverages Fourier Transforms for global information capture and spatial gating units for local feature emphasis. This study’s methodology encompasses a sophisticated augmentation and patching strategy at the input level, followed by a series of MLP blocks designed to extract hierarchical features and spatial relationships, culminating in a global average pooling layer before classification. Evaluated against state-of-the-art MLP-based models including MLP-Mixer, FNet, gMLP, and ResMLP across diverse and extensive CT datasets, including abdominal, and chest scans, DiagnosticMLP demonstrated a remarkable ability to converge efficiently, with competitive accuracy, F1 scores, and AUC metrics. Notably, in datasets featuring kidney and abdomen disorders, the model showcased superior generalization capabilities, underpinned by its unique design that addresses the complexity inherent in CT imaging. The findings in terms of accuracy and precision-recall balance posit DiagnosticMLP as an exceptional outperforming alternative to attention-reliant models, paving the way for streamlined, efficient, and scalable AI tools in medical diagnostics, reinforcing the potential for AI-augmented precision medicine without the dependency on attention-based architectures. Full article

Robust visual place recognition (VPR) enables mobile robots to identify previously visited locations. For this purpose, the extracted visual information and place matching method plays a significant role. In this paper, we critically review the existing VPR methods and group them into three major categories based on visual information used, i.e., handcrafted features, deep features, and semantics. Focusing the benefits of convolutional neural networks (CNNs) and semantics, and limitations of existing research, we propose a robust appearance-based place recognition method, termed SVS-VPR, which is implemented as a hierarchical model consisting of two major components: global scene-based and local feature-based matching. The global scene semantics are extracted and compared with pre-visited images to filter the match candidates while reducing the search space and computational cost. The local feature-based matching involves the extraction of robust local features from CNN possessing invariant properties against environmental conditions and a place matching method utilizing semantic, visual, and spatial information. SVS-VPR is evaluated on publicly available benchmark datasets using true positive detection rate, recall at 100% precision, and area under the curve. Experimental findings demonstrate that SVS-VPR surpasses several state-of-the-art deep learning-based methods, boosting robustness against significant changes in viewpoint and appearance while maintaining efficient matching time performance. Full article

Food-image recognition plays a pivotal role in intelligent nutrition management, and lightweight recognition methods based on deep learning are crucial for enabling mobile deployment. This capability empowers individuals to effectively manage their daily diet and nutrition using devices such as smartphones. In this study, we propose an Efficient Hybrid Food Recognition Net (EHFR–Net), a novel neural network that integrates Convolutional Neural Networks (CNN) and Vision Transformer (ViT). We find that in the context of food-image recognition tasks, while ViT demonstrates superiority in extracting global information, its approach of disregarding the initial spatial information hampers its efficacy. Therefore, we designed a ViT method termed Location-Preserving Vision Transformer (LP–ViT), which retains positional information during the global information extraction process. To ensure the lightweight nature of the model, we employ an inverted residual block on the CNN side to extract local features. Global and local features are seamlessly integrated by directly summing and concatenating the outputs from the convolutional and ViT structures, resulting in the creation of a unified Hybrid Block (HBlock) in a coherent manner. Moreover, we optimize the hierarchical layout of EHFR–Net to accommodate the unique characteristics of HBlock, effectively reducing the model size. Our extensive experiments on three well-known food image-recognition datasets demonstrate the superiority of our approach. For instance, on the ETHZ Food–101 dataset, our method achieves an outstanding recognition accuracy of 90.7%, which is 3.5% higher than the state-of-the-art ViT-based lightweight network MobileViTv2 (87.2%), which has an equivalent number of parameters and calculations. Full article

For the synthesis of ultra-large scene and ultra-high resolution videos, in order to obtain high-quality large-scene videos, high-quality video stitching and fusion are achieved through multi-scale unstructured array cameras. This paper proposes a network model image feature point extraction algorithm based on symmetric auto-encoding and scale feature fusion. By using the principle of symmetric auto-encoding, the hierarchical restoration of image feature location information is incorporated into the corresponding scale feature, along with deep separable convolution image feature extraction, which not only improves the performance of feature point detection but also significantly reduces the computational complexity of the network model. Based on the calculated high-precision feature point pairing information, a new image localization method is proposed based on area ratio and homography matrix scaling, which improves the speed and accuracy of the array camera image scale alignment and positioning, realizes high-definition perception of local details in large scenes, and obtains clearer synthesis effects of large scenes and high-quality stitched images. The experimental results show that the feature point extraction algorithm proposed in this paper has been experimentally compared with four typical algorithms using the HPatches dataset. The performance of feature point detection has been improved by an average of 4.9%, the performance of homography estimation has been improved by an average of 2.5%, the amount of computation has been reduced by 18%, the number of network model parameters has been reduced by 47%, and the synthesis of billion-pixel videos has been achieved, demonstrating practicality and robustness. Full article

Context: Despite the effort put into developing standards for structuring construction costs and the strong interest in the field, most construction companies still perform the process of data gathering and processing manually. This provokes inconsistencies, different criteria when classifying, misclassifications, and the process becomes very time-consuming, particularly in large projects. Additionally, the lack of standardization makes cost estimation and comparison tasks very difficult. Objective: The aim of this work was to create a method to extract and organize construction cost and quantity data into a consistent format and structure to enable rapid and reliable digital comparison of the content. Methods: The approach consisted of a two-step method: firstly, the system implemented data mining to review the input document and determine how it was structured based on the position, format, sequence, and content of descriptive and quantitative data. Secondly, the extracted data were processed and classified with a combination of data science and experts’ knowledge to fit a common format. Results: A large variety of information coming from real historical projects was successfully extracted and processed into a common format with 97.5% accuracy using a subset of 5770 assets located on 18 different files, building a solid base for analysis and comparison. Conclusions: A robust and accurate method was developed for extracting hierarchical project cost data to a common machine-readable format to enable rapid and reliable comparison and benchmarking. Full article

The evaluation of the level of farmland infrastructure is a necessary objective condition for the use of arable land and agricultural development. In order to investigate the evaluation index system and method of farmland infrastructure level, this article uses the Kenli District of the Yellow River Delta as the research region. In the study region, six typical observation sample areas are chosen. Each area receives high-resolution UAV photos, which are then used to extract information about the farmland infrastructure of the field. A farming infrastructure evaluation index system was built, consisting of 10 indexes for four aspects, including farmland roads, field plots, ditches, and forest belts, using the 100 m by 100 m grid method to divide the evaluation units. The comprehensive index technique was used to calculate the farmland infrastructure score of each unit and identify the degree of excellent, good, and poor farmland infrastructure. The weight of each indication was decided by the hierarchical analysis method. There were 20 excellent grades, 77 good grades, and 29 poor grades among the 126 evaluation units in the study area, with excellent and good grades accounting for 79.13% and area proportions of 14.29%, 64.84%, and 20.87%, respectively. Among the six sample areas, sample areas E and F had the highest percentages of excellent grades, sample area A had 82.62% of the good grades, and all sample areas except A and C had a percentage of poor grades that was higher than 20%. Regularity of the fields, average size of the fields, and the agricultural plot’s slope are the dominant indexes of farmland infrastructure in each observation sample area, and the indexes of the ratio of the perimeter of roads to the perimeter of fields, density of ditches, and the ratio of area of agricultural forest networks to area of fields need to be optimized and improved. The spatial distribution of each grade differs significantly. The evaluation results are consistent with the real situation in the study region and have positive reference meaning for the development and management of farming infrastructure, according to the study’s proposed evaluation system and methodology. Full article

Skin lesion detection is crucial in diagnosing and managing dermatological conditions. In this study, we developed and demonstrated the potential applicability of a novel mixed-scale dense convolution, self-attention mechanism, hierarchical feature fusion, and attention-based contextual information technique (MSHA) model for skin lesion detection using digital skin images of chickenpox and shingles lesions. The model adopts a combination of unique architectural designs, such as a mixed-scale dense convolution layer, self-attention mechanism, hierarchical feature fusion, and attention-based contextual information, enabling the MSHA model to capture and extract relevant features more effectively for chickenpox and shingles lesion classification. We also implemented an effective training strategy to enhance a better capacity to learn and represent the relevant features in the skin lesion images. We evaluated the performance of the novel model in comparison to state-of-the-art models, including ResNet50, VGG16, VGG19, InceptionV3, and ViT. The results indicated that the MSHA model outperformed the other models with accuracy and loss of 95.0% and 0.104, respectively. Furthermore, it exhibited superior performance in terms of true-positive and true-negative rates while maintaining low-false positive and false-negative rates. The MSHA model’s success can be attributed to its unique architectural design, effective training strategy, and better capacity to learn and represent the relevant features in skin lesion images. The study underscores the potential of the MSHA model as a valuable tool for the accurate and reliable detection of chickenpox and shingles lesions, which can aid in timely diagnosis and appropriate treatment planning for dermatological conditions. Full article

A multispectral infrared zoom optical system design and a single-frame hierarchical guided filtering image enhancement algorithm are proposed to address the technical problems of low contrast, blurred edges, and weak signal strength of single-spectrum infrared imaging of faint targets, which are easily drowned out by noise. The multispectral infrared zoom optical system, based on the theory of complex achromatic and mechanical positive group compensation, can simultaneously acquire multispectral image information for faint targets. The single-frame hierarchical guided filtering image enhancement algorithm, which extracts the background features and detailed features of faint targets in a hierarchical manner and then weights fusion, effectively enhances the target and suppresses the interference of complex background and noise. Solving multi-frame processing increases data storage and real-time challenges. The experimental verification of the optical system design and image enhancement algorithm proposed in this paper separately verified that the experimental enhancement was significant, with the combined use improving Mean Square Error (MSE) by 14.32, Signal-Noise Ratio (SNR) by 11.64, Peak Signal-to-Noise Ratio (PSNR) by 12.78, and Structural Similarity (SSIM) by 14.0% compared to guided filtering. This research lays the theoretical foundation for the research of infrared detection and tracking technology for clusters of faint targets. Full article

The metro station ridership features are associated significantly with the built environment factors of the pedestrian catchment area surrounding metro stations. The existing studies have focused on the impact on total ridership at metro stations, ignoring the impact on varying patterns of metro station ridership. Therefore, the reasonable identification of metro station categories and built environment factors affecting the varying patterns of ridership in different categories of stations is very important for metro construction. In this study, we developed a data-driven framework to examine the relationship between varying patterns of metro station ridership and built environment factors in these areas. By leveraging smart card data, we extracted the dynamic characteristics of ridership and utilized hierarchical clustering and K-means clustering to identify diverse patterns of metro station ridership, and we finally identified six main ridership patterns. We then developed a newly built environment measurement framework and adopted multinomial logistic regression analysis to explore the association between ridership patterns and built environment factors. (1) The clustering analysis results revealed that six station types were classified based on varying patterns of passenger flow, representing distinct functional characteristics. (2) The regression analysis indicated that diversity, density, and location factors were significantly associated with most station function types, while destination accessibility was only positively associated with employment-oriented type stations, and centrality was only associated with employment-oriented hybrid type station. The research results could inform the spatial planning and design around metro stations and the planning and design of metro systems. The built environment of pedestrian catchment areas surrounding metro stations can be enhanced through rational land use planning and the appropriate allocation of urban infrastructure and public service facilities. Full article

The mainstream methods for change detection in synthetic-aperture radar (SAR) images use difference images to define the initial change regions. However, methods can suffer from semantic collapse, which makes it difficult to determine semantic information about the changes. In this paper, we proposed a hierarchical fusion SAR image change-detection model based on hierarchical fusion conditional random field (HF-CRF). This model introduces multimodal difference images and constructs the fusion energy potential function using dynamic convolutional neural networks and sliding window entropy information. By using an iterative convergence process, the proposed method was able to accurately detect the change-detection regions. We designed a dynamic region convolutional semantic segmentation network with a two-branch structure (D-DRUNet) to accomplish feature fusion and the segmentation of multimodal difference images. The proposed network adopts a dual encoder–single decoder structure where the baseline is the UNet network that utilizes dynamic convolution kernels. D-DRUNet extracts multimodal difference features and completes semantic-level fusion. The Sobel operator is introduced to strengthen the multimodal difference-image boundary information and construct the dynamic fusion pairwise potential function, based on local boundary entropy. Finally, the final change result is stabilized by iterative convergence of the CRF energy potential function. Experimental results demonstrate that the proposed method outperforms existing methods in terms of the overall number of detection errors, and reduces the occurrence of false positives. Full article

Visual geo-localization plays a crucial role in positioning and navigation for unmanned aerial vehicles, whose goal is to match the same geographic target from different views. This is a challenging task due to the drastic variations in different viewpoints and appearances. Previous methods have been focused on mining features inside the images. However, they underestimated the influence of external elements and the interaction of various representations. Inspired by multimodal and bilinear pooling, we proposed a pioneering feature fusion network (MBF) to address these inherent differences between drone and satellite views. We observe that UAV’s status, such as flight height, leads to changes in the size of image field of view. In addition, local parts of the target scene act a role of importance in extracting discriminative features. Therefore, we present two approaches to exploit those priors. The first module is to add status information to network by transforming them into word embeddings. Note that they concatenate with image embeddings in Transformer block to learn status-aware features. Then, global and local part feature maps from the same viewpoint are correlated and reinforced by hierarchical bilinear pooling (HBP) to improve the robustness of feature representation. By the above approaches, we achieve more discriminative deep representations facilitating the geo-localization more effectively. Our experiments on existing benchmark datasets show significant performance boosting, reaching the new state-of-the-art result. Remarkably, the recall@1 accuracy achieves 89.05% in drone localization task and 93.15% in drone navigation task in University-1652, and shows strong robustness at different flight heights in the SUES-200 dataset. Full article